JPH1184438A - Crosscorrelator - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a correlator having the measuring function of a self- correlation signal. SOLUTION: This crosscorrelation meter measures the signal light to be measured as a function of a variable delay by applying the variable delay to sampling light and combining this light with the signal light to be measured, then focusing this light to making the same incident on a medium having an optical nonlinear effect and converting the nonlinear signal generated in this medium to an electric signal. A branching mirror 114 which branches the sampling light and a synthesizing mirror 118 which aligns the optical axis of the one sampling light which is branched by this branching mirror 114 and does not give the delay to the optical axis of the signal light to be measured are added to the meter described above.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

[0001] The present invention relates to a cross-correlator. More specifically, the present invention relates to a technique for measuring an intensity waveform of an ultrafast optical signal in a picosecond to femtosecond region. In particular, the present invention provides a device having a function to accurately and easily obtain the time resolution related to waveform measurement.

[0002]

2. Description of the Related Art In recent years, the use of ultrahigh-speed optical signals in the picosecond to femtosecond range has been prosperous in many industrial and academic fields including optical communication and optical information processing. There has been an increasing demand for devices for measuring intensity waveforms in detail. Here, as a method of measuring the intensity waveform, a method of first converting a light signal to be measured into a similar electric signal using a high-speed photodetector and observing the waveform with existing electric signal measuring means is considered first. .

However, in this case, it is necessary that the photodetector used has a shorter response time than the optical signal to be measured. In the ultra-short time region from the picosecond to the femtosecond region, there is no photodetector that satisfies this condition.

[0004] Therefore, in the ultra-short time range, the intensity waveform measurement by the cross-correlator described below must be used. The cross-correlator uses an optical pulse that is synchronized with the optical signal to be measured and has a shorter time width than the optical signal to be measured, and crosses this optical pulse and the optical signal to be measured in a medium exhibiting an optical nonlinear effect. The magnitude of the generated nonlinear signal
It is measured as a function of the time relationship between the signal light to be measured and the light pulse.

[0005] This is because a signal light to be measured is gated by a short light pulse, and the gated signal output is measured while changing the time position of the signal light to be measured at the gate, thereby obtaining a signal light waveform to be measured. Is measured, and it is a method similar to sampling measurement for an electric signal. Therefore, the measurement using the cross-correlator is referred to as optical sampling, and the optical pulse with a short time width used in conjunction is referred to as a sampling optical pulse. There are various orders and types of optical nonlinear effects used, and consequently the manner in which nonlinear signals are taken.

For example, as the second-order optical nonlinear effect, a sum frequency generation effect or a two-photon absorption effect is often used. In the former, inverse of the sum of the wavelength of the incident two light, i.e. a wavelength lambda _1, lambda respect _{λ 2 1 λ 2 / (λ} 1 +
Since light (sum frequency light) having a wavelength of [lambda] ₂ ) is generated, the light is converted by a photodetector having sensitivity to this sum frequency light wavelength to obtain an electric signal proportional to the power of the sum frequency light.

[0007] In the latter, the transmittance of the sampling light pulse decreases with the two-photon transition that occurs when one photon is obtained from each of the two incident lights, so that the photodetector having sensitivity to the wavelength of the sampling light pulse. To obtain an electrical signal proportional to the change in transmittance. Alternatively, in the case of a semiconductor material, an electric signal proportional to the two-photon transition probability can be directly obtained by attaching an electrode and collecting carriers generated by the two-light transition.

As the third-order optical nonlinear effect, an optical Kerr effect or a coupling tone generating effect is used, and the signal is converted by an appropriate photodetector to obtain an electric signal proportional to the magnitude of each effect. The time resolution related to the optical sampling measurement is determined by the response time of the optical nonlinear medium or element used and the time width of the sampling light pulse, and does not depend at all on the response time of a means for converting an electric signal such as a photodetector.

This property is the reason why optical sampling measurement using a cross-correlator is performed in a very short time range from the picosecond to the femtosecond range. Further, non-resonant non-linear effects without a real one-photon transition are typically used, in which case the optical non-linear medium or element can be considered to exhibit a practically instantaneous response.

Therefore, it can be considered that the time resolution of the optical sampling measurement by the cross-correlation meter is usually determined only by the time width of the sampling optical pulse. A signal obtained when the sampling optical pulse is subjected to optical sampling measurement by itself is particularly called an autocorrelation signal.
If the autocorrelation signal is measured in addition to the cross-correlation signal with the optical signal to be measured, the time resolution based on the time width of the sampling optical pulse can generally be estimated.

Conventionally, a device having an optical delay line obtained by modifying a Michelson-type interferometer has been used as a cross-correlator. FIG. 6 shows this conventional example. This device uses Applie
The autocorrelator published in d Physics Letters, Vol. 27 (1975), pp. 488-490 is changed to a cross-correlator, and furthermore an optical fiber incident type is adopted.

In this example, the sampling light pulse is transmitted from the sampling optical fiber 603 to the fiber connector 6.
After being incident through the optical axis 04 and converted into parallel rays by the collimating lens 605, it reaches the right-angle reflecting mirror 626 via the right-angle reflecting mirror 629 and the right-angle reflecting prism 631. The sampling light pulse further enters the reflector 627, and after being turned back, reaches the focusing lens 621.
On the other hand, the signal light to be measured enters from the signal optical fiber 606 via the fiber connector 607,
08, it is converted into a parallel light beam,
0, and reaches the reflector 628 via the right-angle reflecting prism 631.

The signal light to be measured is further transmitted to the reflector 62
After being turned back by 8, the light reaches the converging lens 621 via the right-angle reflecting mirror 613. Here, as the reflectors 627 and 628, a configuration in which two plane mirrors are bonded so as to be orthogonal to each other or a so-called corner reflector configuration in which three plane mirrors are bonded so as to be orthogonal to each other is used. These two reflectors determine the length of the optical path of the sampling light pulse from the reflection point on the right-angle reflecting prism 631 to the focusing lens 621 via the right-angle reflecting mirror 626 and the reflector 627, and the optical signal under measurement. From a reflection point on the right-angle reflecting prism 631 through a reflector 628 and a right-angle reflecting mirror 613 to form a focusing lens 621.
The optical path lengths up to the top are arranged so as to be equal to each other.

Further, the reflector 627 on the sampling light pulse side is disposed on the parallel moving base 611, and the reflector 627 is arranged around a position where the lengths of the two optical paths are equal to each other.
It is configured to be able to translate in a direction parallel to the optical axis of the light incident on the light source 27. The sampling light pulse and the signal light to be measured having passed through the focusing lens 621 cross each other at the same time in the nonlinear crystal 622 placed at the focal point of the focusing lens 621.

Thus, the signal light to be measured and the sampling light pulse interact with each other in the nonlinear crystal with a high light intensity. In this example, the sum frequency generation effect in the non-linear crystal is used as the non-linear effect. In this case, the generated sum frequency light is emitted in a direction sandwiched by two incident lights.

The stop 623 is arranged so as to have an opening in the emission direction of the sum frequency light. As a result, only the sum frequency light generated by the interaction between the signal light to be measured and the sampling light pulse is transmitted to the stop 623. Through the condenser lens 6
24 can be reached. This sum frequency light is applied to the photodetector 6
25, an electric signal proportional to the magnitude is obtained.

Here, the parallel moving table 611 is driven, and while the reflector 627 on the sampling light pulse side is translated, the output electric signal of the photodetector 625 is measured and recorded. Here, when the reflector 627 retreats by the distance x, the time at which the sampling light pulse enters the nonlinear crystal 622 is delayed by 2x / c from the time at which the signal light to be measured enters.

Here, c represents the speed of light, and the first multiplier 2 reflects that the sampling light pulse is turned back on the reflector 627. For example, for x of 0.15 mm, a delay of 1.0 ps occurs. Thus, the function of performing optical sampling measurement is realized in this conventional cross-correlator.

[0019]

However, the above-mentioned conventional cross-correlation apparatus has a problem that it is difficult to measure an autocorrelation signal easily. As described above, in order to estimate the time resolution based on the time width of the sampling light pulse,
It is necessary to measure the autocorrelation signal.

When the sampling light pulse is of an optical fiber incidence type as in the conventional apparatus, the necessity for measuring the autocorrelation signal is particularly increased. The pulse width of the sampling light pulse is generally shorter than that of the signal light to be measured, and even if the signal light to be measured in the picosecond region is handled, the sampling light pulse having a width from the subpicosecond to the femtosecond region is used. It is usually used.

Such an ultrashort-duration light pulse is liable to be deformed by the chromatic dispersion of the optical fiber. Therefore, even if the pulse width immediately after the light source that generated the light pulse is known, the sampling light There is no guarantee that the original pulse width will be maintained when propagating and exiting the fiber. Therefore, it is necessary to measure the autocorrelation signal of the optical pulse actually emitted from the sampling optical fiber and estimate the time resolution based on the width.

Further, it is known that the time resolution itself due to the time width of the sampling light pulse can be relaxed if the autocorrelation signal is measured precisely as one proceeds. Let us see this in accordance with the present example using the sum frequency generation effect. When writing the intensity waveform of the sampling light pulse as i ₀ (t) and the intensity waveform of the signal light under measurement as i ₁ (t) as a function of time t, the cross-correlation signal g _c (τ) is expressed by the following equation (1). ).

[0023]

(Equation 1)

Here, τ is the difference between the time at which the sampling light pulse enters the nonlinear crystal and the time at which the signal light to be measured enters, ie, the delay time, and the integrand i ₁ (t) i
₀ (t−τ) is the intensity waveform of the sum frequency generated by the nonlinear crystal, and the integral represents the effect of photoelectric conversion by the photodetector having a sufficiently slow response time compared to the sampling light pulse. On the other hand, the autocorrelation signal g ₀ of the sampling light pulse
(Τ) is represented by the following equation (2).

[0025]

(Equation 2)

Here, considering the Fourier transform of the delay time τ of these correlation signals, the cross-correlation signal is calculated by the following equation (3), and the auto-correlation signal is calculated by the following equation (4) ). G _c (ω) = I ₁ (ω) I ₀ ^* (ω) (3) G ₀ (ω) = I ₀ (ω) I ₀ ^* (ω) (4)

[0027] _{Here, I 0 (ω), I} 1 (ω) , respectively, the intensity of the sampling optical pulse waveform i ₀ (t), is the Fourier transform of the intensity of the measured signal light waveform i ₁ (t) .
Now, assuming that the waveform of the sampling light pulse is symmetrical, I ₀ (ω) is a real number, and I ₀ (ω) is obtained by taking the square root of both sides in equation (4). By substituting this into equation (3), I ₁ (ω) becomes the following equation (5)
Is calculated by I ₁ (ω) = G _c (ω) / (G ₀ (ω)) ^1/2 (5)

By subjecting the obtained I ₁ (ω) to inverse Fourier transform, the intensity waveform of the signal light to be measured, i ₁ (t)
Can be requested. Thus, in principle,
The autocorrelation signal is used to remove the influence of the time width of the sampling light pulse from the cross-correlation signal, and the cross-correlation signal as if using an infinitely thin sampling light pulse, in other words, the intensity waveform of the signal light to be measured , And this operation is called decorrelation operation.

In general, it is not guaranteed that the sampling light pulse is strictly symmetrical. However, it is known that the pulse generated by the current femtosecond light source is highly symmetrical. Even after propagation through the optical fiber, the symmetric property is preserved as long as deformation due to second-order chromatic dispersion (group velocity dispersion) of the optical fiber is considered. Thus, the above decorrelation operation can be performed with considerable reliability in practice.

From a different viewpoint, for a sampling light pulse having a very short time width, the measurement of the auto-correlation signal is inevitable in order to obtain the pulse width at the light source in the first place. Therefore, if the cross-correlation device is provided with an auto-correlation signal measuring function, there is no need to use a separate auto-correlation device, and there is an advantage that the equipment can be reduced in cost and size. This is an expected effect regardless of whether the sampling light pulse is of the optical fiber incidence type.

As described above, the autocorrelation signal is a signal obtained by optically sampling and measuring the sampling light pulse itself in place of the signal light to be measured. It may seem that all of the light in fiber 603 should be extracted and incident on signal optical fiber 606.

FIG. 7 shows two optical fibers which are connected to a fiber type 3 dB coupler based on such an idea, and are branched and emitted 1: 1.
The correlation signals collected and connected to the fiber connectors 604 and 607, respectively, are shown. With respect to the sampling light pulse emitted from the sampling optical fiber in this case, the width of the correct autocorrelation waveform obtained by the dedicated autocorrelator was 200 fs.

However, the width of the correlation signal shown in FIG.
00fs, and it can be seen that the autocorrelation waveform is not correct at first glance. This is the fiber type 3dB used
This is because the wavelength dispersion of the coupler is large, and the light pulse having a very short time width is greatly deformed.

Further, the correlation signal shown in FIG. 7 has a slight left-right asymmetry. In general, the autocorrelation signal always has the property of being symmetrical about the delay time.
In this sense, the signal shown in FIG. 7 does not satisfy the requirement of being an autocorrelation signal. This result is due to the fact that the wavelength dispersion of both branches of the used coupler is non-uniform, and the light pulse undergoes different deformations in each branch.

Further, the signal shown in FIG. 7 is not located around the origin of the delay time and appears with a shift of about 10 ps. Here, the origin of the delay time of the cross-correlator is set so that the correlation signal becomes maximum when pulses are simultaneously incident on the two fiber connectors 604 and 607.

Accordingly, the displacement of the signal in FIG.
The length of both output fibers of the fiber coupler used is non-uniform about 2 mm, which suggests that the side connected to the fiber connector 607 on the signal optical fiber side is long. The uniformity of the length of the fiber adhering to the fiber type coupler is not normally subject to quality control as a specification, and it is difficult to avoid such nonuniformity in length.

However, when two fibers from the fiber coupler are connected to the correlator, it is very inconvenient in practice that the delay time position where the signal appears cannot be predicted.

As described above, in order to easily perform autocorrelation measurement using a conventional cross-correlation device having no autocorrelation measurement function, a sampling optical pulse is distributed and supplied externally using a fiber type coupler. The strategy is
(1) The chromatic dispersion of the fiber coupler cannot be ignored.
In addition, (2) there is a problem that the delay time position where the autocorrelation signal appears is shifted from the origin of the delay time of the cross-correlator, which is hardly a fundamental solution.

According to this, the sampling optical fiber 603, the fiber connector 604,
What about a device that removes the collimating lens 605, the signal optical fiber 606, the fiber connector 607, and the collimating lens 608 to make it a spatial beam incident type?

In this case, a semi-transparent mirror is externally inserted on the sampling optical axis to split a part of the sampling light, and a reflecting mirror is inserted on the incident optical axis of the signal light to be measured.
The split light is incident on the right-angle reflecting mirror 630. In this configuration, the chromatic dispersion of the first problem described above can be kept sufficiently small.

However, the problem of the second delay time position is rather severe in this case. Right angle reflecting mirror 629,
Assuming that the reflection point of the semi-transmissive mirror and the reflecting mirror 630 forms a vertex of a rectangle, the optical path on the signal light side has a larger optical axis of the signal light to be measured and the optical axis of the sampling light than the optical path on the sampling light side. It becomes longer by the interval. This is equal to the spacing between the two fiber connectors 604, 607, which in this prior art device is 70 mm.

This corresponds to a delay time of 233 ps.
This exceeds the variable range of the delay time of the conventional device. Therefore, when the conventional apparatus is of a spatial beam incident type, it is impossible to extract an autocorrelation signal by branching a part of the sampling light outside.

If the interval between the two incident beams is made narrower, the shift of the delay time position can be made smaller, but it is apparent that it cannot be reduced to zero in principle. As described above, the conventional cross-correlator does not have a function of measuring an auto-correlation signal. (1) Estimation of time resolution or calculation of decorrelation cannot be performed. (2) Optical branching to the outside. Even if a system is added, there is a problem that it is difficult to collect an autocorrelation signal due to a shift in the delay time position.

An object of the present invention is to solve these difficulties and to provide a correlator having a function of measuring an autocorrelation signal.

[0045]

The cross-correlator of the present invention comprises:
It is characterized in that a splitting mirror for splitting the sampling light and a combining mirror for making the split sampling light coincide with the optical axis of the signal light to be measured are added. The length of the optical path from the incident end of the sampling light through the split mirror to the synthetic mirror,
It is preferable that the length of the optical path from the incident end of the signal light to be measured to the synthetic mirror be equal.

As means for switching between cross-correlation measurement and auto-correlation measurement, the branch mirror is a semi-transparent mirror, the synthetic mirror is a total reflection mirror, and the semi-transparent mirror and the reflection mirror are formed so as to be removable. Alternatively, both the split mirror and the synthetic mirror may be semi-transmissive mirrors so that the cross-correlation signal and the auto-correlation signal are superimposed and collected, and both measurements may be performed at once.

[Operation] A variable delay is given to the sampling light, and it is incident together with the signal light to be measured into a medium exhibiting an optical non-linear effect, and the non-linear signal generated in the medium is converted into an electric signal. In the cross-correlator measuring as a function of the above, a splitting mirror for splitting the sampling light and a combining mirror for making the split sampling light coincide with the optical axis of the signal light to be measured are added. As a result, the sampling light branched instead of the signal under measurement is subjected to optical sampling measurement, and the function of measuring the autocorrelation signal is provided.

At this time, the length of the optical path from the incident end of the sampling light to the combining mirror via the branching mirror and the length of the optical path from the incidence end of the signal light to be measured to the combining mirror are:
If they are formed to be equal, the origin of the delay time related to the cross-correlation measurement and the origin of the delay time related to the auto-correlation measurement coincide, and it is possible to prevent the shift of the delay time position. If the branch mirror is a semi-transparent mirror, the synthetic mirror is a total reflection mirror, and these semi-transparent mirrors and the total reflection mirror are formed so as to be able to be inserted and removed, it will completely disturb the adjustment of the optical system related to the original cross-correlation measurement. It is possible to switch between cross-correlation measurement and auto-correlation measurement without any need.

Alternatively, if both the splitting mirror and the combining mirror are semi-transparent mirrors, the cross-correlation signal and the auto-correlation signal are superimposed and sampled, and simultaneous measurement of both signals is realized. Even for a correlator having an optical delay line modified from the Michelson-type interferometer shown in the conventional example, a branching mirror for branching the sampling light and a branching sampling light for the signal light to be measured are formally formed. It is possible to add a synthetic mirror that matches the axis.

For example, in FIG.
If the split mirror is inserted into the optical path from the mirror mirror 627 to the reflecting mirror 627, the combining mirror is inserted into the optical path from the right-angle reflecting prism 631 to the reflector 628, and the reflected light from the split mirror is incident on the combining mirror by the right-angle reflecting mirror. Good. However, according to this, it is possible to make the length of the optical path from the incident end of the sampling light to the combining mirror through the branch mirror equal to the length of the optical path from the incidence end of the signal light to be measured to the combining mirror. I can never do that.

That is, the optical path configuration obtained by modifying the Michelson-type interferometer as shown in the conventional example cannot exert the sufficient effect of the present invention. Therefore, in the correlator according to the present invention, the optical path from the incident end of the sampling light to the focusing lens to the nonlinear crystal and the optical path from the incident end of the signal light to be measured to the focusing lens are key-shaped. Here, the three sides of these key shapes are sequentially described as a shaft portion, a horizontal portion,
I will call it the key part.

With the key type of the sampling optical path and the key type of the signal light path to be measured, each axis, each horizontal part, and each key are configured to be parallel to each other. Further, the total length of the key type of the sampling optical path and the key type of the signal light path to be measured are equal to each other, and the key-type horizontal part of the signal light path to be measured is connected to the key-type horizontal part of the sampling light path. It is configured to be longer than.

At this time, the key-shaped horizontal portion of the optical path of the signal under test and the key-shaped axis of the sampling optical path intersect. Here, on the key-shaped shaft of the sampling optical path,
A branching mirror is arranged at a point between the end point and the intersection, and a composite mirror is arranged at a bending point to an axis and a middle point of the intersection on a key-shaped horizontal group of the signal light path to be measured.

In the reflecting mirror connecting the splitting mirror and the combining mirror, the optical path from the splitting mirror to the reflecting mirror is parallel to the key-shaped horizontal portion, and the optical path from the reflecting mirror to the combining mirror is parallel to the key-shaped shaft. Arrange so that At this time, the optical path length from the combining mirror to the right-angle reflecting mirror becomes permanently equal to the optical path length from the bending point to the shaft to the combining mirror.

As a result, the length of the optical path from the incident end of the sampling light to the combining mirror via the split mirror and the length of the optical path from the incidence end of the signal light to be measured to the combining mirror are:
Become equal. The above discussion holds regardless of the angle between the two key-shaped shafts and the horizontal part, or the angle between the horizontal part and the key part. However, the variable delay without changing the optical axis of the key part in the sampling light. In order to provide the following, it is required that the two key-shaped shaft portions and the key portion are parallel to each other.

[0056]

Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing a basic configuration of a cross-correlator of the present invention, and shows a state where both a sampling light and a signal light to be measured are of a spatial beam incident type. In this configuration, the portion of the cross-correlator relating to the acquisition of the nonlinear signal is the same as the cross-correlator of the conventional example, and the focusing lens 121,
It comprises a nonlinear crystal 122, a stop 123, a condenser lens 124, and a photodetector 125.

Here, as in the conventional example, the case where the sum frequency generation effect in the nonlinear crystal 122 is used as the nonlinear effect is illustrated. In the illustrated example, the key-shaped optical path of the sampling light is transmitted from the sampling light incident end 101 through the right-angle reflecting mirror 109 and further through the right-angle reflecting mirror 110 to the focusing lens 1.
It reaches 21.

The key-shaped optical path of the signal light to be measured extends from the signal light incident end 102 to the converging lens 121 via the right angle reflecting mirror 112 and the right angle reflecting mirror 113. The length of the axis of the keyed optical path of the sampling light, that is, the sampling light incident end 1
Let α be the distance from 01 to the right-angle reflector 109, and h be the length of the horizontal part, ie, the distance from the right-angle reflector 109 to the right-angle reflector 110.

On the other hand, the length of the axis portion of the key-shaped optical path of the signal light to be measured, that is, the distance from the signal light incident end 102 to the right-angle reflecting mirror 112 is α ', and the length of the horizontal portion, that is, the right-angle reflecting mirror 112
Is written as h '. The condition that the total length of the key type of the sampling optical path and the total length of the key type of the signal light path to be measured are equal is expressed by the following equation (6). h′−h = 2 (α−α ′) (6)

Here, the difference between the lengths of the key portions in the two key types is equal to the difference between the lengths of the shaft portions. In the figure, the axis of the key-shaped optical path of the sampling light and the key-shaped horizontal part of the optical path of the signal under test intersect at the optical path intersection 126.

On the sampling optical path, this optical path intersection 1
Before reaching 26, the branch mirror 114 is arranged. Also, on the optical path of the signal under test, the right-angle reflecting mirror 112 and the optical path intersection 12
The combining mirror 118 is arranged at the bisecting point of No. 6.

A right angle reflecting mirror 116 is disposed at the intersection of a straight line drawn in parallel with the horizontal portion of the key-shaped optical path from the branch mirror 114 and a straight line drawn in parallel with the axis of the key-shaped optical path from the combining mirror 118. As a result, the length of the optical path from the sampling light incidence end 101 to the combining mirror 118 via the branch mirror 114 and the right angle reflection mirror 116, and the length of the optical path from the signal light incidence end 102 to the right angle reflection mirror 11
The length of the optical path that reaches the combining mirror 118 via the line 2 becomes equal.

This is because the sum of the distance from the sampling light incident end 101 to the split mirror 114 and the distance from the right angle reflecting mirror 116 to the combining mirror 118 are equal to the distance from the signal light incident end 102 to the right angle reflecting mirror 112. And the branch mirror 11
The distance from the right-angle reflector 116 to the right-angle reflector 116 is
It will be readily appreciated that this is equal to the distance from 2 to the synthetic mirror 118.

On the optical path side of the sampling light, the right angle reflecting mirror 10
9 and the right-angle reflecting mirror 110 are arranged on the translation stage 111, and the optical axis incident on the right-angle reflecting mirror 109, that is, the key shape of the sampling optical path, around the position where the equation (6) holds, that is, around the origin of the delay time. Is configured to be capable of translating in a direction parallel to the axis of. With this translation, the key-shaped key portion of the sampling optical path, that is, the output optical axis from the right-angle reflector 110 to the focusing lens 121 is not displaced, so that the incident optical axis to the right-angle reflector 109 and the right-angle reflector 110 Need to be parallel.

Here, the parallel moving table 111 is driven to measure and record the electric signal output from the photodetector 125 while translating the right-angle mirror 119 and the right-angle mirror 110 on the sampling light pulse side. Thus, in the cross-correlator of this configuration, a function of performing optical sampling measurement is realized.

At this time, a cross-correlation signal equivalent to that of the conventional cross-correlator is obtained with the split mirror 114 and the combining mirror 118 removed. On the other hand, the sampling light split by the splitting mirror passes through the autocorrelation optical path 117 and is combined with the composite mirror 11.
In the state where the signal 8 flows into the signal light path, an autocorrelation signal of the sampling light is obtained.

Further, at this time, the autocorrelation signal always takes the maximum value at the position where the equation (6) holds, that is, at the origin of the delay time. Thus, in the cross-correlator of the present invention, the origin of the delay time for the cross-correlation measurement always coincides with the origin of the delay time for the auto-correlation measurement.

[0068]

【Example】

Embodiment 1 FIG. 2 shows a first embodiment of the present invention. In this embodiment, in order to switch between the cross-correlation measurement and the auto-correlation measurement, both the split mirror and the synthetic mirror can be inserted and removed.

FIG. 2 shows a configuration in which both the sampling light and the signal light to be measured are of an optical fiber incidence type. As shown in FIG. 2, the sampling optical fiber 1 is connected to the sampling light incident end via a fiber connector 104.
The sampling light pulse emitted from the fiber at the tip of 03 is converted into a parallel spatial beam by the collimating lens 105.

The distal end of the signal optical fiber 106 is located at the signal light incident end via the fiber connector 107,
The measured optical signal emitted from the fiber is converted into a parallel spatial beam by the collimating lens 108. In this case, as two collimating lenses 105 and 108,
If the one having uniform wavelength dispersion characteristics is used, the key-shaped optical path of the sampling light from the fiber connector 104 to the right angle reflecting mirror 109 and further to the converging lens 121 via the right angle reflecting mirror 110 and the right angle from the fiber connector 107 according to the equation (6). The total length of the key-shaped optical paths of the signal light to be measured which reaches the focusing lens 121 via the reflecting mirror 112 and the right-angle reflecting mirror 113 is made equal, and the origin of the delay time can be set.

For example, the axis length α is set to 155 mm and the horizontal length h is set to 70 mm for the keyed optical path of the sampling light, and the axis length α ′ is set to 118 for the keyed optical path of the signal light to be measured.
mm, and the horizontal length h ′ is set to 144 mm.
Holds, and the total lengths of both key type optical paths are equal. As the branch mirror, θ = 45 ° incident 1: 1 branch semi-transparent mirror 115 obtained by depositing a dielectric multilayer film on a quartz glass substrate having a thickness of d = 1 mm
Was used.

It is recommended that the substrate of the semi-transparent mirror be thin for two reasons. The first is the viewpoint of the wavelength dispersion of the substrate, where n is the refractive index of the substrate and D is the dispersion per unit propagation length.
⁽⁰⁾ , the chromatic dispersion D when transmitted through the substrate
Is proportional to the thickness d as shown in the following equation (7).

D = nD ⁽⁰⁾ d / (n ² −sin ² θ) ^1/2 (7) In the equation (7), the value at a wavelength of 1.55 μm of quartz glass, that is, n = 1.444, D ⁽⁰⁾ = -28 fs ² / mm
Is substituted, −32 fs ² is obtained as the wavelength dispersion D of the semi-transparent mirror 115 of this embodiment. The spread of the pulse width due to the chromatic dispersion was only 0.06% even for a pulse of 50 fs, and it was possible to suppress the chromatic dispersion to a sufficiently small value.

The second is the viewpoint of the lateral displacement of the optical axis due to the substrate, and the lateral displacement amount w of the output beam before and after the insertion of the substrate.
Is also proportional to the thickness d, as shown in the following equation (8). w = d ｛1−cos θ / (n ² −sin ² θ) ^1/2 ｝ sin θ (8) From equation (8), the displacement amount w associated with the semi-transparent mirror 115 of the present embodiment.
Can be estimated to be 0.3 mm.

This lateral displacement is caused by two right-angle reflecting mirrors 109,
After passing through 110, it appears twice on the converging lens 121. In the case of the present embodiment, the distance between the sampling light and the signal light to be measured on the focusing lens 121 is 4 mm.
The beam diameter of both light beams was set to 1.5 mm.

It can be said that a lateral displacement of 0.6 mm from this state is within a marginally allowable range. As a synthetic mirror, a right angle reflecting mirror 119 was used. The position of the right-angle reflecting mirror 119 was set at a point on the key-shaped horizontal portion of the optical path of the signal light to be measured, which was exactly half the distance between the two fiber connectors 104 and 107 from the right-angle reflecting mirror 112.

Since the distance between the right angle reflecting mirror 112 and the optical path intersection 126 is equal to the distance between the two fiber connectors 104 and 107, the distance between the right angle reflecting mirror 112 and the optical path intersection 126 is divided into two by the combining mirror 118. Will be satisfied. In the case of this embodiment, the distance between the two fiber connectors 104 and 107 is 70 mm.
Therefore, the position of the right angle reflecting mirror 119 was set to a point of 35 mm measured from the right angle reflecting mirror 112.

In this embodiment, the semi-transparent mirror 115 and the right-angle reflecting mirror 119 are mounted on switchable mounts, respectively, and can be separated from the optical path. This type of mount has a function of accurately reproducing the original azimuth and elevation when the mounted mirror is returned to the optical path.

In the case of the present embodiment, a mount is used which switches between insertion and detachment by rotating the mirror together with an azimuth / elevation angle adjustment mechanism about a rotation axis below the mirror mounting portion. Of course, it goes without saying that other types of mounts can be applied in the same manner, for example, by switching the insertion / removal by sliding the mirror, or by switching the mirror in the plane together with the adjustment mechanism to switch the insertion / removal.

In the apparatus of the present embodiment, when measuring the autocorrelation, both the semi-transparent mirror 115 and the right-angle reflecting mirror 119 are inserted into the optical path. At this time, the signal light to be measured entering from the signal optical fiber 106 is blocked by the right-angle reflecting mirror 119 and does not transmit toward the right-angle reflecting mirror 113.

Only the sampling light branched by the semi-transparent mirror 115 and following the autocorrelation light path 117 is reflected by the right-angle reflection port 1.
9 and enters the right angle reflecting mirror 113. Here, by driving the translation stage 111 and translating the right-angle reflecting mirror 109 and the right-angle reflecting mirror 110 on the side of the sampling light pulse, the output electric signal of the photodetector 125 is measured and recorded. A correlation signal as shown in FIG. 3 (a) was obtained.

Since this is a result of optical sampling measurement of the sampling light itself, it corresponds to an autocorrelation signal.
This autocorrelation signal appears around the origin of the delay time, and the width of the autocorrelation waveform is a correct value of 200 fs.

The sampling light pulse is a Cr-doped Y
A pulse having a center wavelength of 1.55 μm generated by an AG (yttrium aluminum garnet) laser. The width of the autocorrelation signal measured immediately after the laser light source is 1
30 fs. The reason why the width of the autocorrelation waveform shown in FIG. 3A is slightly widened is that the pulse spread caused by the chromatic dispersion of the optical fiber accompanying propagation through the sampling optical fiber is reflected. I have.

On the other hand, in the apparatus of this embodiment, at the time of cross-correlation measurement, both the semi-transparent mirror 115 and the right-angle reflecting mirror 119 mounted on the switchable mount are separated from the optical path. As a result, both are shown in FIG.
1 and the position shown by the right-angle reflecting mirror 119-1 and are removed from the respective optical paths. At this time,
The signal light to be measured entering from the signal optical fiber 106 enters the right angle reflecting mirror 113. Since the semi-transparent mirror 115 has been removed, the sampling light is not split and nothing appears on the autocorrelation optical path 117.

Here, it is not a principle requirement that the semi-transparent mirror 115 be separated during the cross-correlation measurement, and the semi-transparent mirror 115 is not required.
The cross-correlation measurement can be performed even with the insertion. Semi-transparent mirror 11
In the case where 5 is left inserted, it is not necessary to consider the above-described lateral displacement of the optical axis due to the substrate.

However, if the semi-transparent mirror 115 is left inserted, there is a problem that the sampling light for branching 1: 1 and following the autocorrelation optical path is wasted, and the sensitivity of the cross-correlation measurement is reduced by half. . Therefore, at the time of the cross-correlation measurement, it is better that the semi-transparent mirror 115 is also separated from the optical path.

Here, while driving the translation stage 111 to translate and move the right-angle mirror 109 and the right-angle mirror 110 on the sampling light pulse side, the output electric signal of the photodetector 125 is measured and recorded. Thus, a correlation signal as shown in FIG. 3B was obtained. Since this is a result of optical sampling measurement of the signal light to be measured, it corresponds to a cross-correlation signal.

In this cross-correlation signal, a sharp change comparable to the auto-correlation signal shown in FIG. 3A is not found, so that in the case of this measurement, the details of the waveform of the signal light to be measured are grasped. It can be estimated that sufficient time resolution is obtained. As described above, the apparatus of the present embodiment was able to easily switch between the auto-correlation measurement and the cross-correlation measurement while maintaining the origin of the delay time.

[Embodiment 2] FIG. 4 shows a second embodiment of the present invention.
Shown in In this embodiment, since the transflective mirror is used for both the split mirror and the synthetic mirror, the cross-correlation measurement and the auto-correlation measurement can be performed simultaneously. In this case, since the semi-transparent mirror 115 at the branch mirror position and the semi-transparent mirror 120 at the composite mirror position are fixed at the inserted positions, the switchable mount used in the above embodiment is unnecessary.

As these semi-transparent mirrors, a dielectric multilayer film was deposited on a quartz glass substrate having a thickness of d = 1 mm and θ = 45 °
The one having an incident 1: 1 branch was used. The substrate of this semi-transparent mirror is
It is recommended to be thin from the viewpoint of the wavelength dispersion of the substrate described above. Here, it is not necessary to consider the lateral displacement of the optical axis due to the substrate of the semi-transparent mirror.

In the apparatus of this embodiment, the sampling light pulse incident from the sampling optical fiber 103 is branched 1: 1 by the semi-transparent mirror 115, and half of the pulse reaches the semi-transparent mirror 120 along the autocorrelation optical path 117. Here, the reached sampling light pulse is branched into 1: 1, and half of the pulse goes to the right angle reflecting mirror 113.

As a result, in the autocorrelation measurement, one fourth of the original sampling light pulse is sampled and measured by one half of the sampling light pulse. On the other hand, the signal light to be measured incident from the signal optical fiber 106 is split 1: 1 by the semi-transmissive mirror 120, and half of the light reaches the right-angle reflecting mirror 113.

Therefore, as for the cross-correlation measurement, half of the signal light to be measured is sampled and measured by a half of the sampling light pulse. From this, the measurement sensitivity in the apparatus of the present embodiment is one-half for the autocorrelation measurement and one-fourth for the cross-correlation measurement as compared with the first embodiment. Others are the same as in the first embodiment.

In the apparatus of the present embodiment, the parallel moving table 111 is driven to translate and move the right-angle mirror 109 and the right-angle mirror 110 on the sampling light pulse side while measuring and recording the output electric signal of the photodetector 125. As a result, a correlation signal as shown in FIG. 5 was obtained. On this correlation signal, an autocorrelation signal around the origin of the delay time and a cross-correlation signal around the delay time of 3.3 ps are weighed and measured at once.

At the time of this measurement, in order to prevent the auto-correlation signal and the cross-correlation signal from being overlapped around the origin of the delay time and becoming inseparable, a signal path length of 1 mm was previously added to the signal light to be measured outside the apparatus of this embodiment. Was given a delay. Further, in order to make the wave heights of the autocorrelation signal and the cross-correlation signal comparable, the peak intensity of the sampling light emitted from the sampling optical fiber 103 and the peak intensity of the signal light to be measured emitted from the signal optical fiber are measured. The strength is adjusted outside the device of this embodiment so that the ratio becomes approximately 2: 1.

As described above, depending on the apparatus of this embodiment,
The autocorrelation measurement and the cross-correlation measurement could be performed simultaneously while keeping the origin of the delay time.

[0097]

As described in detail above, according to the cross-correlator of the present invention, the measurement of the auto-correlation signal of the sampling light pulse is improved regardless of whether the light beam is of the spatial beam incidence type or the optical fiber incidence type. Accuracy is achieved, and the position of the delay time where the autocorrelation signal appears is fixed around the origin of the delay time of the cross-correlator, and the adjustment of the optical system related to the original cross-correlation measurement is not disturbed at all. Therefore, when the cross-correlator is used, the autocorrelation signal of the sampling light pulse can be easily measured, and the time resolution can be estimated based on the width of the sampling light pulse or the decorrelation operation can be performed. The present invention can be widely used irrespective of the type of optical nonlinear medium or element used in the correlator, and regardless of the spatial beam incidence type or the optical fiber incidence type, and can improve the accuracy of the cross-correlation measurement. A great effect is obtained industrially.

[Brief description of the drawings]

FIG. 1 is an explanatory diagram showing a basic configuration of a cross-correlator of the present invention.

FIG. 2 is an explanatory diagram showing a first embodiment of the cross-correlator of the present invention.

FIGS. 3A and 3B are graphs showing measurement results according to the first example of the present invention, wherein FIG. 3A shows a measurement result of autocorrelation of sampling light, and FIG. 3B shows a measurement result of cross-correlation.

FIG. 4 is an explanatory diagram showing a second embodiment of the cross-correlator according to the present invention.

FIG. 5 is a graph showing measurement results according to a second example of the present invention.

FIG. 6 is an explanatory diagram showing a configuration of a conventional cross-correlator.

FIG. 7 is a graph showing a result of autocorrelation measurement by a conventional apparatus.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 101 Sampling light incidence end 102 Signal light incidence end 103 Sampling optical fiber 104 Fiber connector 105 Collimating lens 106 Signal optical fiber 107 Fiber connector 108 Collimating lens 109, 110 Right angle reflecting mirror 111 Parallel moving stand 112, 113 Right angle reflecting mirror 114 Branch mirror 115 Semi-transparent mirror 115-1 Semi-transparent mirror (during departure) 116 Right angle reflecting mirror 117 Autocorrelation optical path 118 Synthetic mirror 119 Right-angle reflecting mirror 119-1 Right-angle reflecting mirror (during departure) 120 Semi-transparent mirror 121 Focusing lens 122 Non-linear crystal 123 Stop 124 Condenser Lens 125 Photodetector 126 Optical path intersection 603 Sampling optical fiber 604 Fiber connector 605 Collimating lens 606 Signal optical fiber 607 Fiber connector 608 Collimating lens 6 11 Parallel translation table 613 Right angle reflecting mirror 621 Converging lens 622 Nonlinear crystal 623 Stop 624 Condensing lens 625 Photodetector 626 Right angle reflecting mirror 627,628 Reflector 629,630 Right angle reflecting mirror 631 Right angle reflecting prism

Claims (4)

[Claims] 1. A variable delay is applied to a sampling light, and after being combined with a signal light to be measured, the light is incident and focused on a medium having an optical nonlinear effect, and a nonlinear signal generated in the medium is converted into an electric signal. In a cross-correlator for measuring a signal light to be measured as a function of the variable delay, a splitting mirror for splitting the sampling light, and an optical axis of one of the sampling lights split by the splitting mirror and giving no delay, A cross-correlator characterized by adding a synthetic mirror that matches the optical axis of light. 2. The optical system according to claim 1, wherein a length of an optical path from the incident end of the sampling light to the combining mirror via the branch mirror is equal to a length of an optical path from the incidence end of the signal light to be measured to the combining mirror. The cross-correlator according to claim 1, wherein 3. The apparatus according to claim 2, wherein the split mirror is a semi-transmissive mirror, the composite mirror is a total reflection mirror, and the semi-transparent mirror and the total reflection mirror are respectively detachable from an optical path. Item 7. The cross-correlator according to Item 1. 4. The cross-correlator according to claim 1, wherein both the split mirror and the synthetic mirror are semi-transmissive mirrors.